Hyperspectral image sensors help characterize polymers in electronic waste to improve recycling

Plastics make up around a quarter of the materials contained in electronic waste (e-waste). The proportion that is recycled is comparatively low—the majority is simply incinerated. The first step to improve recycling is the identification of polymer materials, so that they can be selectively sorted and processed in a way that preserves their function.

Researchers at the Helmholtz Institute Freiberg for Resource Technology (HIF), an institute of the Helmholtz-Zentrum Dresden-Rossendorf (HZDR), have now succeeded in determining the specific characterization of the main e-waste plastic types by combining multiple sensors. Applied on an industrial scale, more plastics can be optimally processed and returned to the production chain.

The work is published in the journal Waste Management.

Almost all electronic devices contain plastics, also known as polymers. These polymers are specialized for certain functions. The aim is to recycle them in such a way that they can be reused for equivalent applications. So first, they must be identified according to their composition.

Sorting them according to type is a major challenge for recycling companies, particularly due to the high proportion of black polymers in e-waste. The roughly shredded e-waste ends up on conveyor belts in the recyclers’ sorting plants and is scanned by infrared sensors. Black plastics are not recognized, as the color black absorbs the wavelengths covered by the infrared sensor.

As a result, black plastics in particular are often thermally recycled, which means incinerated. Another problem is downcycling, a deterioration in the quality of the recycled waste compared to the original material. A successful recycling process must ensure that the polymer-specific functionalities are retained in order to enable reuse with consistent quality.

Scientists at the HIF examined 23 polymers using imaging and point-measurement spectral sensors, identifying the decisive parameters for reliable and robust differentiation of plastic types. The high speed at which the polymers move on the conveyor belt poses an additional challenge. The sensors must therefore detect and characterize the components quickly in order to find the optimal way for further processing.

“In order to evaluate the performance potential of the sensors, they must be used under the operating conditions that prevail in recycling plants. At the HIF, we have a conveyor belt test track on which the materials move at speeds of up to one meter per second and are sequentially scanned by multiple sensors,” HIF scientist Dr. Andréa de Lima Ribeiro explains the test procedure.

It all depends on the right combination

The scientists worked with hyperspectral image sensors (HSI), which capture image data with several hundred color channels. Raman spectroscopy is also used, in which the material is irradiated with a laser to generate material-specific light scattering.

The resulting spectrum allows conclusions to be drawn about the material under investigation. Furthermore, a FTIR spectrometer (Fourier transform infrared spectrometer) was used, with high spectral resolution and wide detection range.

The FTIR detection range was supplemented by a high-resolution spectroradiometer in the visible to short-wave infrared range. Both manual point sensors not only validated the results from the imaging sensors, but also provided additional information regarding plastic composition, particularly for black plastics.

“The investigation has shown that none of the sensors alone is able to identify all plastic types and at the same time meet the industry operational requirements. The results demonstrate the good suitability of HSI sensors for the specific identification of transparent and light-colored plastic types,” de Lima Ribeiro said.

“Raman spectroscopy enabled the point identification of all polymer types, including black plastics. The experiments also show the successful identification of plastics even at short acquisition times of 500 milliseconds. The optimal characterization of the plastics is achieved with the combination of imaging and point measurements.”

Process is already being used in the recycling of automotive parts

Sensor-based plastic characterization is already being used in the Car2Car project, in which the HIF is involved. The aim of the project is to develop automated material detecting concepts for the most important material groups in cars (steel, aluminum, glass, plastic and copper) in order to improve the separation and processing of these secondary raw materials by type.

“Metals and plastics are often closely interlinked in end-of-life products. We have therefore further developed the sensor technology so that it can distinguish metals and polymers from one another and differentiate between types relevant to the process. This is essential for the reuse of the raw materials contained in end-of-life vehicles,” explains Dr. Margret Fuchs, scientist in the field of optical sensors and sensor systems at the HIF.

The application of the specific sensors is based on the results of the RAMSES-4-CE research project, in which multi-sensor systems for the rapid identification of critical compounds were investigated in terms of their performance and accuracy.